Realism's Emotional Edge in AI-Generated Pet Food Ads: Mediating Effects on Purchase Intention
(A2026-134269)

AUTHORS

Thomas Reichstein, BTU Cottbus-Senftenberg; Florian Dost, Brandenburg University of Technology

ABSTRACT

Advertising research highlights the role of visual-evoked emotions in shaping consumer intentions beyond cognition. With AI enabling rapid generation of realistic or stylized images, a key question arises: Does visual realism outperform abstraction in eliciting positive emotions that boost purchase intention? Grounded in Construal Level Theory (CLT), which links concrete visuals to low-level construal and intensified affect, versus abstract ones to high-level construal and muted responses, this study examines positive emotions as a mediator. In a survey experiment (N=131), participants evaluated AI-generated pet food ads with varying abstraction levels (realistic vs. comic-like). Structural equation modeling supported partial mediation: Lower abstraction enhanced positive emotions, which in turn increased purchase intention. Results advance CLT applications in AI-driven branding and advise practitioners to favor realism for emotional engagement in pet ads.